Quantitative liver function tests are based on the clearance concept and measure the plasma disappearance of a test compound such as galactose. Metabolism is inferred to be predominantly hepatic, and usually no knowledge is obtained of the true time course of metabolite formation. Dynamic 31phosphorus magnetic resonance spectroscopy after intravenous administration of fructose directly measures hepatic sugar metabolism. To determine the feasability and the utility of 31P magnetic resonance spectroscopy, we studied the responses of six healthy subjects and nine patients with nonalcoholic cirrhosis to a fructose load. Results were related to the impairment of hepatic function assessed by the galactose-elimination capacity test. Liver spectra were acquired in a 1.5 T whole-body nuclear magnetic resonance unit with a surface coil (9-cm diameter) placed ventrally on the liver; the one-dimensional chemical-shift imaging technique was used to obtain spectra from tissue slices parallel to the surface coil. After a basal spectrum had been obtained, fructose (250 mg/kg) was injected intravenously, and further spectra were collected sequentially every 6 min for 1 hr. Formation of monophosphate esters (9% ± 5% vs. 20% ± 8% of total area; p < 0.01) and utilization of inorganic phosphate (5% ± 4% vs. 11% ± 3% of total area; p < 0.005) were markedly decreased in cirrhotic patients. These measures correlated with the severity of the impairment of liver function measured by the galactose-elimination capacity (r = 0.53 to 0.69; p < 0.05). We conclude that dynamic 31P magnetic resonance spectroscopy is a safe, clinically feasible test that allows detailed insights into biochemical events in liver disease. (HEPATOLOGY 1992;15:835–842).